Feature: Create Result class to store all info about pipeline and results

[cmalinmayor]

In order to implement #95, we need to have access to all info about matcher params and metric params along with the eventual metric output and any other metadata needed to have a complete record of what was done.

We should implement a Result object that is returned from Metric.compute() with the follwing info:

• Matcher info: a dictionary containing the matcher name and any parameters, e.g. {”name”: IOU Matcher, “threshold”: 0.5}
• Metrics info with name and parameters, similar to matcher
• Results (generic dictionary with whatever the metric wants)
• Traccuracy version (tag and commit)
• Other metadata to help users identify which run their results are from, including the "name" of the solution and gt TrackingGraph (this would require also letting users provide a name for their TrackingGraph), or date/time of run

We need the Matcher info to also be stored in the Matched object for continuity of information.
Then #95 is simply a to_dict and/or to_file function in this class. run_metrics can call that automatically with a user-provided or default location with one metric for file or all metrics in one file. Details TBD 🙂